def execute(self, sentence, ident=None):
"""
Convert sentence to speech, preprocessing out unsupported ssml
The method caches results if possible using the hash of the
sentence.
Args:
sentence: Sentence to be spoken
ident: Id reference to current interaction
"""
sentence = self.validate_ssml(sentence)
create_signal("isSpeaking")
if self.phonetic_spelling:
for word in re.findall(r"[\w']+", sentence):
if word.lower() in self.spellings:
sentence = sentence.replace(word,
self.spellings[word.lower()])
key = str(hashlib.md5(sentence.encode('utf-8', 'ignore')).hexdigest())
wav_file = os.path.join(owo.util.get_cache_directory("tts"),
key + '.' + self.audio_ext)
if os.path.exists(wav_file):
LOG.debug("TTS cache hit")
phonemes = self.load_phonemes(key)
else:
wav_file, phonemes = self.get_tts(sentence, wav_file)
if phonemes:
self.save_phonemes(key, phonemes)
vis = self.visime(phonemes)
self.queue.put((self.audio_ext, wav_file, vis, ident))
def test_wait_while_speaking(self):
# Check that test terminates
create_signal('isSpeaking')
Thread(target=wait_while_speaking_thread).start()
sleep(2)
self.assertFalse(done_waiting)
check_for_signal('isSpeaking')
sleep(2)
self.assertTrue(done_waiting)
def test_check_signal(self):
if exists('/tmp/OwO'):
rmtree('/tmp/OwO')
# check that signal is not found if file does not exist
self.assertFalse(check_for_signal('test_signal'))
# Check that the signal is found when created
create_signal('test_signal')
self.assertTrue(check_for_signal('test_signal'))
# Check that the signal is removed after use
self.assertFalse(isfile('/tmp/OwO/ipc/signal/test_signal'))
def execute(self, sentence, ident=None):
"""request and play mimic2 wav audio
Args:
sentence (str): sentence to synthesize from mimic2
ident (optional): Defaults to None.
"""
create_signal("isSpeaking")
sentence = self._normalized_numbers(sentence)
chunks = sentence_chunker(sentence, self.chunk_size)
for idx, req in enumerate(self._requests(chunks)):
results = req.result().json()
audio = base64.b64decode(results['audio_base64'])
vis = self.visime(results['visimes'])
key = str(
hashlib.md5(chunks[idx].encode('utf-8', 'ignore')).hexdigest())
wav_file = os.path.join(get_cache_directory("tts"),
key + '.' + self.audio_ext)
with open(wav_file, 'wb') as f:
f.write(audio)
self.queue.put((self.audio_ext, wav_file, vis, ident))
def _start_listener(message):
"""
Force OwO to start listening (as if 'Hey OwO' was spoken)
"""
create_signal('startListening')
def test_is_speaking(self):
create_signal('isSpeaking')
self.assertTrue(owo.audio.is_speaking())
# Check that the signal hasn't been removed
self.assertTrue(check_for_signal('isSpeaking'))
self.assertFalse(owo.audio.is_speaking())
def test_create_signal(self):
create_signal('test_signal')
self.assertTrue(isfile('/tmp/OwO/ipc/signal/test_signal'))
def process(self, data):
# TODO: Look into removing this emit altogether.
# We need to check if any other serial bus messages
# are handled by other parts of the code
if "owo.stop" not in data:
self.bus.emit(Message(data))
if "Command: system.version" in data:
# This happens in response to the "system.version" message
# sent during the construction of Enclosure()
self.bus.emit(Message("enclosure.started"))
if "owo.stop" in data:
if has_been_paired():
create_signal('buttonPress')
self.bus.emit(Message("owo.stop"))
if "volume.up" in data:
self.bus.emit(Message("owo.volume.increase",
{'play_sound': True}))
if "volume.down" in data:
self.bus.emit(Message("owo.volume.decrease",
{'play_sound': True}))
if "system.test.begin" in data:
self.bus.emit(Message('recognizer_loop:sleep'))
if "system.test.end" in data:
self.bus.emit(Message('recognizer_loop:wake_up'))
if "mic.test" in data:
mixer = Mixer()
prev_vol = mixer.getvolume()[0]
mixer.setvolume(35)
self.bus.emit(Message("speak", {
'utterance': "I am testing one two three"}))
time.sleep(0.5) # Prevents recording the loud button press
record("/tmp/test.wav", 3.0)
mixer.setvolume(prev_vol)
play_wav("/tmp/test.wav").communicate()
# Test audio muting on arduino
subprocess.call('speaker-test -P 10 -l 0 -s 1', shell=True)
if "unit.shutdown" in data:
# Eyes to soft gray on shutdown
self.bus.emit(Message("enclosure.eyes.color",
{'r': 70, 'g': 65, 'b': 69}))
self.bus.emit(
Message("enclosure.eyes.timedspin",
{'length': 12000}))
self.bus.emit(Message("enclosure.mouth.reset"))
time.sleep(0.5) # give the system time to pass the message
self.bus.emit(Message("system.shutdown"))
if "unit.reboot" in data:
# Eyes to soft gray on reboot
self.bus.emit(Message("enclosure.eyes.color",
{'r': 70, 'g': 65, 'b': 69}))
self.bus.emit(Message("enclosure.eyes.spin"))
self.bus.emit(Message("enclosure.mouth.reset"))
time.sleep(0.5) # give the system time to pass the message
self.bus.emit(Message("system.reboot"))
if "unit.setwifi" in data:
self.bus.emit(Message("system.wifi.setup", {'lang': self.lang}))
if "unit.factory-reset" in data:
self.bus.emit(Message("speak", {
'utterance': owo.dialog.get("reset to factory defaults")}))
subprocess.call(
'rm ~/.OwO/identity/identity2.json',
shell=True)
self.bus.emit(Message("system.wifi.reset"))
self.bus.emit(Message("system.ssh.disable"))
wait_while_speaking()
self.bus.emit(Message("enclosure.mouth.reset"))
self.bus.emit(Message("enclosure.eyes.spin"))
self.bus.emit(Message("enclosure.mouth.reset"))
time.sleep(5) # give the system time to process all messages
self.bus.emit(Message("system.reboot"))
if "unit.enable-ssh" in data:
# This is handled by the wifi client
self.bus.emit(Message("system.ssh.enable"))
self.bus.emit(Message("speak", {
'utterance': owo.dialog.get("ssh enabled")}))
if "unit.disable-ssh" in data:
# This is handled by the wifi client
self.bus.emit(Message("system.ssh.disable"))
self.bus.emit(Message("speak", {
'utterance': owo.dialog.get("ssh disabled")}))
if "unit.enable-learning" in data or "unit.disable-learning" in data:
enable = 'enable' in data
word = 'enabled' if enable else 'disabled'
LOG.info("Setting opt_in to: " + word)
new_config = {'opt_in': enable}
user_config = LocalConf(USER_CONFIG)
user_config.merge(new_config)
user_config.store()
self.bus.emit(Message("speak", {
'utterance': owo.dialog.get("learning " + word)}))